Evolving coherent and non-trivial biclusters from gene expression data: An evolutionary approach

Biclustering in microarray data is used to discover a set of genes expressed similarly in a subset of conditions. Biclustering algorithms require to identify coherent and non-trivial biclusters, i.e., the biclusters should have low mean squared residue and high row variance. This article presents a genetic algorithm based biclustering technique that optimizes a combination of these objectives. A novel encoding strategy is proposed. The performance of the proposed algorithm has been evaluated on two benchmark real life gene expression data sets and compared with some other well-known biclustering techniques.